Speaker
Description
The Belle II experiment at the SuperKEKB collider in Japan is a next-generation B-factory with a large international collaboration and demanding computing needs. Belle II has been using Rucio as its data management system since early 2021, supporting global distribution and access to physics data. The experiment is now transitioning to Rucio as its primary metadata service, tightly integrating it with DIRAC. Some recent developments include the evaluation of TimeScaleDB for efficient tracking of dataset popularity to guide data placement, the integration of Large Language Models to provide a natural interface for querying and managing data, and the deployment of OIDC tokens as a replacement for X.509 certificates. We will present operational experience, performance results, and lessons learned that may inform similar efforts in other experiments.
